Upper Volta - Rural Roads Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 5234 PROJECT PERFORMANCE AUDIT REPORT UPPER VOLTA RURAL ROADS PROJECT (CREDIT 579-UV) August 17, 1984 Operations Evaluation Department This documnent has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authoriation. COUNTRY EXCHANGE RATES Appraisal year average 1975 1 US$ = CFAF 225 Inteevening year average 1976 240 1977 220 1978 220 1979 220 1980 210 Completion year average 1981 260 Final Disbursement year 1982 (Estimate) 330 Average for Project Implementation Period: 240 ABBREVIATIONS AND ACRONYMS BPP - Bureau of Planning and Programming DPW - Directorate of Public Works DRF - Drought Relief Fund (IDA CREDIT 422-UV) ITC - Interministerial Technical Committee OPEV - Office for Promotion of Voltaic Industry ORD - Regional Development Organization RDF - Rural Development Fund (IDA CREDIT 317-UV) SERN - National Road Maintenance Service SERS - Secondary Road Maintenance Service UNDP - United Nations Development Program FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT UPPER VOLTA RURAL ROADS PROJECT (CREDIT 579-UV) TABLE OF CONTENTS Page No. Preface ........................................................... i Basic Data Sheet.................................................. ii Highlights ........................................................ iv PROJECT PERFORMANCE AUDIT MEMORANDUM I. PROJECT PREPARATION AND BACKGROUND ...... ...... 1 II. PROJECT IMPLEMENTATION .................................. 5 III. POINTS OF SPECIAL INTEREST ....................... .... 8 IV. CONCLUSIONS .............. ............. 11 PROJECT COMPLETION REPORT - FIRST HIGHWAY PROJECT I. Introduction ................................ 15 II. Project Preparation and Appraisal ................. . 17 III. Pr.jJect Implementation ................................... 20 IV. Institutional Development ............................... 33 V. Economic Reevaluation .................................... 35 VI. Association Performance ... ............. .... 36 VII. Conclusions . ........ ........ 36 Tables: 1. Design Standards: At Appraisal and Actual ............. 38 2. Rural Roads Constructed by Region and Year. ........ 39 3. Net of Regional and Rural Roads Maintained by Region and Year .................... ...... 40 4. Road Impact Sftdy: Number of Zones, Road Sections, Villages and Markets Selected for Study ... 42 5. Appraisal Estimate and Actual Project Cost ......... 43 6. Estimation of Average per Kilometer Cost of Roads Improved Under the Project ..... . ........ 44 7. Estimated and Actual Schedule of Disbursements ........ 45 8. Estimated and Actual Allocation of Credit Proceeds ....... 46 Map IBRD 11504 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. PROJECT PERFORMANCE AUDIT REPORT UPPER VOLTA: RURAL ROADS PROJECT (CREDIT 579-UV) PREFACE This is the performance audit report on the Upper Volta Rural Roads Project (Credit 579-UV) of US$7.5 million presented to the Board on June 16, 1975. The Credit was closed in December 1981 and funds were fully disbursed with the exception of US$5,800 which were cancelled. This document consists of a Project Completion Report (PCR) prepared by the West Africa Regional Office, and of a Project Performance Audit Memorandum, prepared by the Opera- tions Evaluation Department (OED). OED has studied the project documentation in the Records Center and has reviewed the PCR against the Appraisal and President's Reports, the legal documents, and the transcripts of the Executive Directors' meeting which con- sidered the project. Discussions were held with Bank operational staff and with consultants who worked on various project components. An OED mission visited Upper Volta in November-December 1983, had discussions with Govern- ment officials and with donor agency representatives resident in Ouagadougou, and travelled on project roads. The PCR summarizes adequately the experience under the project and the audit elaborates on certain aspects of project design and implementation as well as on the importance of close donor-agency-Government collaboration. The draft audit report was sent to the Borrower for comments; how- ever, none were received. PROJECT PERFORMANCE AUDIT BASIC DATA SHEET UPPER VOLTA RURAL ROADS PROJECT (CREDIT 579-UV) KEY PROJECT DATA Appraisal Actual or Item Expectation Current Estimate Total Project Cost (US$ million) 8.5 8.6/a Overrun (%) - 10% Credit Amount (US$ million) 7.5 7.5 Disbursed - 7.494 Cancelled - 0.006 Date Physical Components Completed 06/30/79 12/31/81 Proportion Completed by: Appraisal Completion Date (%) 100 25 Actual Completion Date (Z) - 75-90/b Economic Rate of Return (Z) 10-18 n.a./c OTHER PROJECT DATA Original Actual or Item Plan Est. Actual First Mention in Files or Timetable 12/72 07/02/73 Government's Application Negotiations 06/16/75 05/21-24/75 Board Approval Date 07/15/75 06/16/75 Credit Agreement Date 08/13/75 Effectiveness Date 02/11/76 Closing Date 12/31/79 06/30/81 12/31/81 Borrower Republic of Upper Volta Executing Agency Ministry of Public Works, Transport and Urban Dev. Fiscal Year of Borrower 01/01 - 12/31 Follow-on Project Name Fourth Highway Project Credit Number 1164-UV Amount (US$ million) 46.0 Credit Agreement Date 02/28/82 /a However, only 61% of road kilometrage included at appraisal was actually constructed, although to higher standards, and 88% of road kilometrage originally planned was actually maintained. / Difficult to assess meaningfully due to differences between actual and indicative appraisal program. /c The PCR indicates that a rate of return at the time of project cor- pletion is not available but suggests that for road construction it is 10% or better and for road maintenance it is over 20%. The audit has concluded that absence of traffic data renders the results of any reevaluation debatable. - iii - MISSION DATA Month/ No. of Date of Item Year Weeks Persons Man/Weeks Report Identification 05/73 2 2 4.0 - Preparation 07/74 2 2 4.0 12/18/74 Appraisal 12/74 2 2 4.0 06/06/75 6 12.0 Supervision I 12/75 1.5 1 1.5 01/13/76 Supervision II/a 04/76 1.0 2 2.0 06/03/76 Supervision IfI/a 07/76 1.0 1 1.0 09/24/76 Supervision IV 10/76 1.5 1 1.5 04/15/76 Supervision V 01/77 1.5 1 1.5 02/11/77 Supervision VI/a 04/77 2.0 1 1.0 06/22/77 Supervision VI/a 10/77 2.0 1 2.0 04/30/77 Supervision VIII 02/78 1.0 1 1.0 03/02/78 Supervision IX/b 02/79 1.0 1 1.0 03/05/79 Supervision X 04/79 1.0 1 1.0 05/28/79 Supervision XI/b 06/79 2.0 2 4.0 07/26/79 Supervision XII 10/79 1.5 1 1.5 11/16/79 Supervision XIII 09/80 1.0 1 1.0 04/12/80 Supervision XIV 05/81 1.0 1 1.0 05/22/81 Supervision XV 11/81 1.0 1 1.0 12/04/81 Supervision XVI/c 02/82 2.0 1 1.0 03/19/82 Supervision XVII 03/82 2.0 2 4.0 - 24.0 28.0 /a Combined with supervision of road component of Bougouriba Agricultural Development Project. /b Combined with preparation of Second Rural Road Project. /c Combined with supervision of Fourth Highway Project - iv - PROJECT PERFORMANCE AUDIT REPORT UPPER VOLTA RURAL ROADS PROJECT (CREDIT 579-UV) HIGHLIGHTS Upper Volta has an economy based exclusively on agriculture but productive regions are far from population centers and good transport is crucial for the movement of foodstuffs. Agricultural development of the relatively more productive southern and western areas of the country has been the focus of Bank assistance and most of the Bank-supported agricultural projects implemented until the mid-1970s included rural road improvement components. At the same time, other donors were assisting similar projects but the Government lacked funds and staff to make further improvements on the rural road network and to maintain works already completed. The central objective of the 1975 Rural Roads Project was to staff and equip the Secondary Roads Maintenance Service (SERS) of the Directorate of Public Works (DPW) so that i. could develop and maintain low-cost roads in regions with significant potential for agricultural growth.1/ A separate component was to finance a study of the impact on rural development of roads improved under earlier Bank project. Road Lmprovement and maintenance work was completed with a 30-month delay; the total length of road improvements was 736 k, compared with 1200 km planned at appraisal (para. 10); out of the total 3,300 km planned to be maintained under the project work was carried out on about 3,000 km and, as in the case of improvements, output declined in 1981 (PCR, paras 3.15-3.16). Institution-building at SERS experienced some initial difficulties but by 1979 the organization was beginning to function in a reasonably effective manner. In 1980/81 improvements in SERS operations were largely nullified by a change to a decentralized system of road administration and political com- plications from 1982 on have made course corrections difficult (para. 33). The Impact Study was too ambitious an undertaking and the Government was not particularly interested in it (paras. 12, 24 and 25). Total costs remained within appraisal estimates but costs per kilo- meter of road improved increased some 60%, in part because of greater than anticipated volumes of gravelling (para. 13). The economic reevaluation in 1/ SERS is a misleading name because it implies an exclusive concern with maintenance whereas in fact the responsibilities of this Department also cover new construction and improvement of existing roads. The name was chosen because SERN (National Roads Maintenance Service) already existed although it too was misnamed since it dealt with maintenance as well as new construction and upgrading of national roads. -v - the PCR (para. 5.01) states that an average daily traffic of 10-12 vehicles suffices to yield a return of at least 10% but the audit believes that all reevaluation computations are debatable because of continuing lack of traffic data (para. 15). The audit found that a number of covenants have been breached, in total or in part, and that this is attributable to conditions already prevailing during project preparation which, if they had been taken into due account at the time the project was appraised, would have led to the framing of more realistic covenants during project negotiation (para. 18). The audit found that the SERS institutional capability has been compromised by the shift to decentralization (para 21); that the construction technology used was not, under Voltaic conditions, the most appropriate (paras. 22 and 23); that the project's recurrent cost implications are con- siderable and that it is doubtful whether Government can sustain the burden with its own resources (paras 19 and 20). In sum, the project had sound objectives but neither was the framework of project design quite appropriate for attaining them, nor the indicated time sufficient for SERS to become effective. Recent socio-political developments have added new dimensions which deserve thorough study prior to the formulation of future Bank assis- tance to Upper Volta. Chief among these dimensions is the need for coordina- tion among all aid agencies active in the country (para. 26). - 1 - PROJECT PERFORMANCE AUDIT MEMORANDUM UPPER VOLTA RURAL ROADS PROJECT (CREDIT 579-UV) 1. PROJECT PREPARATION AND BACKGROUND 1. Upper Volta depends on agriculture for the survival of its 6 million inhabitants, over 90% of whom are engaged in subsistence farming. Livestock, cotton and groundnuts are the only exports and the 1980 per capita GNP has been estimated at no more than US$210 equivalent. With one excep- tion,1/ Bank group operations in Upper Volta until 1975 aimed at supporting the development of agriculture.2/ OED has issued a Project Performance Audit Report on the Cotton Roads Project (Report No. 3528 of June 24, 1981) highlighting problems which, throughout the 1970s, plagued the economy and the implementation of development projects. The following pages show that these problems have remained unresolved into the 1980s. 2. The object of the present audit is the Rural Roads Project (Credit 579-UV). It was prepared in 1973 by consultants financed by the Fond d'Aide et de Cooperation (FAC, France); appraised in December 1974; negotiated in May 1975; presented to the Board in June; signed in August; and became effec- tive in February 1976. The project involved a US$7.5 million credit with a total estimated cost of US$ 8.5 million equivalent. It consisted of three parts: the first was aimed at rural road improvements and maintenance; the second at institution-building to strengthen the secondary roads service (SERS) of the Directorate of Public Works (DPW); and the third provided for an evaluation of the impact and effectiveness of the road components included in the agricultural and rural development projects financed by the Bank Group over the previous five years. 3. As early as January 1972, almost two years before appraisal, Bank Group staff had identified factors constraining both the economic development of Upper Volta and the proper design and execution of individual projec.ts. Agriculture was confronted with land desiccation and with a 20% rainfall reduction over twelve years. Public administration suffered from shortages of trained staff and from procedural complexities which slowed down even ordinary transactions. Chronic budgetary shortfalls led to Government unwillingness to allocate funds either for investment or for recurrent expen- ditures as long as there was a chance for donor financing. At the same time, 1/ The First Credit for Telecommunications. 2/ The West Volta Cotton Project (US$6.2 million in 1969); the Rural Development Fund (US$2.2 million in 1972); the Cotton Roads Project (US$2.8 million in 1972, plus a supplementary US$1.35 million in 1974); the Drought Relief Fund (US$2.0 million in 1973) and the Bougouriba Agricultural development Project (US$8.0 million in 1975). -2- no attempt was made by either donors or the Government to rationalize and coordinate the types and flows of foreign aid. Meanwhile, Voltaic migration, primarily to the Ivory Coast, increased as more and more people could not find work at home. 4. In mid-1972, during the preparation of the Cotton Roads Project, the Bank Group suggested the use of labor-based methods for the construction of feeder roads on the grounds that this would create employment and save foreign exchange which otherwise would go for the purchase of equipment, fuel and spares. The Government objected to the suggestion and argued that a switch from equipment-based to labor-based construction technology would require studies for which no funds were available and, also, that labor-based construction would take longer and would probably end up being more costly. The Bank did not insist and the matter was dropped. However, Bank Group con- cern about Government weaknesses in road planning, administration and mainte- nance were sufficiently strong for specific covenants addressing these points to be inserted in the Credit Agreement. The Cotton Roads PCR (paras. 3.17 and 3.18) acknowledges that covenants were too ambitious and that little could be done to enforce them. 5. In view of the preceding considerations, it might have been expected that preparation and appraisal of the Feeder Roads Project over the December 1974-May 1975 period would have taken due account of the organize- tional, staffing, budgetary and donor coordination issues which successive Bank missions had repeatedly reported. This does not appear to have happened. The December 1974 Terms of Reference for the appraisal mission suggest an ambitious project scope; the January 1975 Issues Paper contains no list of critical issues; and the February 1975 Decision Memorandum simply records that assurances will be sought that Government would take appropriate measures to renew construction equipment when needed. No questions were asked about the Government's willingness and ability to raise the funds necessary for equipment replacement although the record had already shown that Government found it easier not to repair but to ask a donor agency for -3- replacements.3/ It is equally noteworthy that the Bank Group which had raised the issue of labor-based construction in 1972, chose not to do so in 1975. Similarly, no questions were raised concerning any of the problem areas which Bank Group staff had been reporting during the previous three years. For example, the Government was expected to contribute US$1.3 million equivalent to the project although Bank Group staff had secured the opinion of well-informed sources that this would not be possible. There was mounting need for donor agency coordination because aid agencies were increasingly operating as self-contained enclaves. The Government was still experiencing staff shortages and this would unavoidably have an adverse effect on project implementation. All of the above notwithstanding, the April 1975 covering note on the draft Appraisal Report stated that no issue of substance exists. Project processing proceeded but staff from different departments of the Bank Group continued, throughout April and May 1975, to express reservattons about aspects of the project which the Appraisal Report presented in too positive a manner. The record shows that staff flatly asserted that Government did not have the means to maintain roads, while other staff expressed doubts as to whether the economic analysis expected to be done by the DPW would in fact be carried out, taking into account the available personnel and the complexity of the methodology proposed for road evaluation. 6. As approved by the Board in June 1975, and as described in the Appraisal Report, the Feeder Roads Project was a complex undertaking whose structure and justification can be summarized as follows: (a) GDP has stagnated since 1970. Agriculture is the main employer but cannot generate enough jobs, and there is massive emigration. The project will support agricultural development (SAR, paras. 2.03 and 2.05; Annex 2). 3/ The Bank's Western Africa Regional Office is of the opinion that the question of Government's willingness and ability to raise funds was not questioned because the Government's preference to replace rather than repair equipment using donor financing seemed to be the only realistic solution given the paucity of mechanical skills in the country. The issue of labor-based construction raised in 1972 had to be abandoned in favor of mechanized construction in 1975 following a pragmatic assessment of what was feasible with the available manpower in Public Works, who were even less able to provide the intensive management inputs needed to organize and implement labor-intensive operations than they were to handle equipment. The general staff shortages of the Government were known, but the creation of a separate organization for feeder road construction and maintenance was expected to mobilize and train local manpower resources. It appears that the project designers seemed convinced that these issues and problems could begin to be tackled through this project, difficult though the task might be, since the alternative would have been an ad-hoc proliferation of feeder road construction by several project entities without any provision for their proper selection in the first instance and for their subsequent maintenance after construction. (b) A number of donors are active in equipment-based rural road con- struction. None of them look after rural road maintenance. The project will help create an institution to correct this (SAR, para. 2.10). (c) Road improvements will have great impact on agricultural produce marketing. Improved transport will result in higher producer prices. The project includes an Impact Study which will generate more specific data (SAR, para. 2.12; Annex 6: Study Terms of Reference). (d) The responsibility for maintaining rural roads had been entrusted to local authorities who proved themselves unequal to the task. Therefore, the Secondary Roads Maintenance Service (SERS) has now been created to take over all tasks related to rural road improve- ment and maintenance (SAR, para. 2.17). (e) The Studies Section (to be established) of SERS will collaborate in matters of road identification, evaluation and preparation with the DPW's Bureau de Planification et de Programmation (BPP) which will henceforth serve as the overall focus for infrastructure planning. The BPP (also to be established) will be largely staffed with ex- patriates paid by the UNDP (SAR, para. 2.18). There are expatri- ates throughout the DPW and one donor agency alone finances the presence of 31 experts (SAR, para. 2.19). (f) The available financial resources for rural roads maintenance do not suffice; consequently, additional resources would be found from budgets of local communities and from self-help contributions (SAR, para. 2.23).4/ (g) There are no Voltaic road contractors. But there are building con- tractors who could become road contractors. The Office for the Promotion of Voltaic Enterprises (Office de Promotion de 1'Enterprise Voltaique, OPEV) will help in this (SAR, paras. 2.24-2.25). (h) The Impact Study (cf. (c) above) provides for 25 man-months of expert services to be commissioned by SERS and would focus on (i) the investigation of the cost-effectiveness of alternative con- struction methods; (ii) analysis of observed traffic; and (iii) linkages between the cost of transport and agricultural production (SAR, para. 3.08). 7. The Board approved the project in June 1975 and the effectiveness date was initially November 1. Time-consuming procedures (cf. para. 3, 4/ The audit found no evidence that rural road maintenance was supported by local communities or self-help contributions. - 5 - above) led the Government to request successive postponements to December 1, January 5, and February 13 to allow more time for legal review. The Bank Group acceded to all three requests, but a December 1975 supervision mission which visited Upper Volta to review status of preparation, reported that pro- ject personnel recruitment was encountering problems and delays, that SEPF was not yet fully established, that the Bank-sponsored Rural Development Fund project staff did not want to turn over to SERS their road construction equipment, that SERS ought to do labor-based construction work but resisted the idea, and that donor agencies seemed quite prepared to continue supplying construction equipment. II. PROJECT IMPLEMENTATION Objectives 8. The project's main purpose was to staff and equip SERS so that it would become capable to develop and maintain low-cost roads in regions with significant potential for agricultural growth. Components 9. The project consisted of three principal components: (a) improvement of about 1,200 km of rural roads and maintenance of about 3,300 km of existing feeder roads; (b) institutional improvements so that DPW could undertake the rational upgrading and maintenance of the rural road network; and (c) evaluation of road components included in earlier Bank-supported projects through an Impact Study to be carried out by consultants. 10. Road improvement and maintenance work was carried out by force account organized in five brigades. Two were set up and equipped as new entities under the project, and three were transferred to SERS from earlier Bank-assisted rural development projects. Construction commenced in 1977 and during the first two years output was about 40 per cent of appraisal esti- mates. First, because equipment of the transferred brigades was about 3 years old and in such poor condition that it took two years to obtain the necessary spares and replacements. Second, because provision of equipment for the new brigades was subject to the usual procurement delays which, in this case, were accentuated by extremely cumbersome Government procedures. Hence, it was no*, until 1979 that outputs began to approach the annual figures targeted at appraisal. The total length of road improvements over a four-year period was 736 km (compared with 1200 km planned at appraisal) but earthworks and gravelling per unit length were more than had been anticipated (PCR, paras. 3.09-3.10). Maintenance operations followed a similar pattern, reaching appraisal output estimates after three years, but the total length taken over for maintenance was about as planned (PCR, paras 3.15-3.16). The -6- audit mission inspected a number of the project roads and found them in generally satisfactory condition. Development of small local contractors did not take place (PCR, para. 3.11). 11. Project assistance to SERS would include financing the costs of three technical experts. It was expected that SERS would make use of BPP in evolving the project work program since, at the time of appraisal, not all roads to be improved had been identified. In its turn, BPP was to receive technical assistance from UNDP. At appraisal it had been envisaged that the work program to be evolved by the BPP and SERS would be reviewed and approved by the Interministerial Technical Committee (ITC) which would ensure that rural road improvements were closely linked with rural development programs (SAR, para. 3.11). In fact, the ITC never functioned as planned and did not function at all until mid-1979 when it approved the 1979/80 and 1980/81 work programs after repeated reminders from the Bank Group (PCR, para. 3.18). The three technical experts to SERS could not be provided as planned and the audit agrees with the PCR (para. 3.05) that this had an adverse effect on the project start, with implications which it was later difficult to overcome. Overall progress was slow but by 1979, three years after project effective- ness, there were signs that SERS was beginning to function in a reasonably effective manner. After 1979, the improvements made up to that point were largely nullified by the change to a decentralized system of road administra- tion which was implemented even though the managerial and technical expertise required to staff the altered organization was only partially available. 12. The Road Impact Study was an ambitious undertaking i-Ltended to overcome the lack of even rudimentary data on rural transport and to improve knowledge of the effects of road access upon agriculture and .the rural econ- omy. Considering the socioeconomic conditions prevailing in Upper Volta and the ground to be covered (SAR, Annex 6: Evaluation Study Terms of Reference), the 25 man-months of expert services budgeted under the project were insufficient. The possibility that the consultants could produce mean- ingful findings was further reduced when Government insisted that surveys be done over the whole country rather than in sample locations (PCR, para. 3.19). Efforts were made to set up a system of regular traffic counts on the main road network and equipment was provided for this purpose but until the December 1983 audit, no counts had been done on rural roads. Such counts as were carried out were done on national roads but were technically deficient because they were undertaken at varying times of the year and it is not pos- sible to extrapolate from them reliable average annual traffic growth rates. Costs and Implementation Period 13. Since the road construction program was indicative only and thus essentially open-ended, financial overexpenditure was very small because work stopped when funds were exhausted (PCR, para. 3.26). The cost per kilometer of road improvement increased some 60% from appraisal estimates (PCR, para. 3.30), an increase not entirely explicable by the greater quantity of gravel used which, in turn, was partly explained by the need for increased earth- works associated with minor structures (PCR, para. 3.10). During project - 7 - execution, the cost per kilometer of road improvements was periodically rees- timated at CFAF 1.1 million (March 1979) and CFAF 1.5 million (November 1980), which include equipment depreciation and overheads but exclude taxes. The reestimated figures are in line with appraisal estimates but significant- ly below the average cost (September 1982) of CFAF 2.3 million shown in the PCR (Table 6), which is roughly 60% higher than the appraisa' estimate (PCR, para. 3.30). The audit considers that the PCR figure is reasonable and was unable to discover why the earlier estimates were not more in line with it. 14. The project was completed with a 30-month delay (PCR, para. 3.25), which contributed to the increase in costs per kilometer. The 10-month delay in project start-up was primarily due to the long time spent in rehabilita- ting the equipment of brigades taken over from other projects. It is note- worthy that the poor condition of this equipment was not detected during appraisal and the delays associated with equipment rehabilitation, plus the delays in works execution, contributed to lowering the project's rate of eco- nomic return. Economic Reevzluation 15. The economic analysis on which the December 1974 appraisal estab- lished the justification of the project (SAR, paras. 4.01-4.06 and Appendix 7) was based on assumptions since there were no data on traffic densities of rural roads (Appendix 7, para. 2). Road user savings originally calculated by consultants appeared excessive to the appraisal mission which scaled them down (Appendix 7, para. 5). The audit has no facts on which to judge the reliability either of the appraisal assumptions or of the summary PCR reesti- mations which have been cs ried out without the benefit of any traffic counts on the project roads (PCR, para. 5.01). In the absence of traffic data it is not possible to say whether the roads are justified since it is not known how many vehicles travel on them. The point is significant because 10 vehicles per day more or less suffice to yield economic rates of return above or below 10%. Covenants 16. Throughout project execution, supervision reports stressed that the covenant requiring the classification of rural roads (Development Credit Agreement, (DCA), Section 4.03), was not fulfilled. Compliance was expected by the end of 1979 but had not been achieved by the time of the December 1983 audit mission. The fact is disturbing because road classification is not only the administrative prerequisite for including roads under the mainte- nance responsibility of DPW, but also provides a basic tool for more effec- tive road planning. 17. Although supervision reports do not bring out any other instances of non-compliance, the audit considers that Government is in breach of a num- ber of other covenants as well. For example, in view of the change to a decentralized administration of the road network in 1981, it cannot be said that the requirements of Section 3.01 of the DCA were followed regarding appropriate administration and employment of competent and sufficient staff. - 8 - Also, although BPP was established under Section 3.02 of the DCA, it had not, up to the time of the December 1983 audit mission, fulfilled its responsibi- lities as defined in sub-paragraph (i) which require that BPP be responsible for the country's entire road network, with respect to collecting and evalu- ating traffic and road inventory data on a permanent basis, formulating and coordinating with all government departments priorities for road improve- ments, analyzing road sector issues, and developing transport policy recom- mendations and road investmqnt plans. Perhaps partly as a consequence of this, ITC did not carry out its duties as laid down in Section 3.04 of the DCA. Finally, questions were raised during the June 1979 supervision mission regarding excessive fuel consumption in the project, indicating thereby a possible non-compliance with DCA Section 3.05(b). 18. The audit considers that many of the above instances of actual or possible non-compliance can be attributed to the political, social and econo- mic conditions in Upper Volta before and during project preparation. How- ever, covenants which proved unenforceable under the Cotton Roads Project (PCR, para. 2.08 (v)) were subsequently repeated almost verbatim under the Rural Roads Project (PCR, para. 2.13 (a)) and this suggests that lessons of experience are not utilized as fully as they could be. III. POINTS OF SPECIAL INTEREST Finance and Management 19. At an early stage of project preparation (January 1972), Bank Group staff put on record that the two basic problems affecting road improvement and maintenance in Upper Volta were shortage of funds and shortage of trained staff. It was noted that equipment provided earlier by FAC had stood idle for want of funds and that any increase in highway activities related either to construction or maintenance would lead to an even greater dependence on expatriates. The project required Government to finance about US$1 million equivalent out of its own resources, and this represented an increase of about 14 per cent in the annual maintenance budget at the time. Total pro- ject financing, including funds provided by the Bank Group but processed through the Government's administrative machinery, would result in a 100 per cent increase in the road maintenance budget, with all the concommitant demands on staff time. In order to finance the eventual replacement of equipment that would be brought in through the project, an extra 25 per cent would have to be added to the road maintenance budget. These were large claims on Government resources but nevertheless the Bank Group was satisfied with a broad assurance that, by the end of the project, Government would pro- vide CFAF 200 million a year for maintenance of rural roads, an increase of 40 per cent over the initial maintenance budget. 20. The net effect of the financial provisions in the DCA would be an immediate doubling of the management burden upon the short-handed highway administration which may possibly have handled the increased work, provided that all the planned technical assistance had materialized on time. In budgetary terms, however necessary the increases noted in the preceding para- graph might have been, the financial straits of the Government made it unlikely that funds would be diverted from other uses. At the same time, and over the identical period during which the Rural Roads Project would be implemented, the managerial and financial burden upon the highway organiza- tion would be further increased by the activities of several other donors implementing highway projects. It is true that most of these activities were carried out almost autonomously under the donor agency concerned and with limited calls on DPW or SERS. Nevertheless, they did make claims upon scarce technical manpower. 21. As a rule, the Bank Group assists institutional and socio-economic growth by operating through Government institutions and by expecting Govern- ments to make financial and management contributions towards the achievement of development objectives. This approach requires a long-term Government commitment to the pursuit of the stated goals even though initial results will be slow to materialize and limited in extent. In the case of the Rural Roads Project, the Bank Group never analyzed the extent and amount of the commitment that Government would have needed to make nor did the Government ever make a conscious commitment. Moreover, results were expected too quick- ly and when some of them started to appear, and the centralized organization started to work, the decision was made by Government and condoned by the Bank, wrongly in the audit's view, to decentralize the administration of highways, although it was questionable whether sufficient resources were available to make this a workable arrangement. Mechanical Equipment 22. The reliance on mechanical equipment for works execution is hard to explain in view of earlier efforts by the Bank to encourage a greater use of casual labor in road construction (cf. para. 4, above). When equipment was transferred to SERS from other projects, it was discovered that with barely half of its expected life gone, it was in such poor state that extensive rehabilitation was needed and repairs took two years to complete. Even aG, and despite the fact that one presumed advantage of machines is greater speed, the thought does not seem to have emerged that equipment-based tech- niques may not be the best method for road construction under Upper Volta conditions and in a country with scarce foreign exchange shortages. Labor- based methods were regarded as suitable for further study (PCR, para. 3.20) but were in fact discarded as an alternative to equipment-based techniques largely on the grounds that they would require major changes in SERS organi- zation. In the audit's view, the argument overlooked that proper management of equipment-based operations require skills which are less likely to be available in Upper Volta than those needed for labor management. The audit found that project equipment, although well administered by workshops, was underutilized due to lack of spares and to a shortage of good operators, both of them chronic phenomena which Government continues to be unable to over- come. It is strange that heavy machinery, representing a capacity equivalent to that of hundreds of laborers, is allowed to stand idle for months at a time without causing much concern, while it is argued that organizing manual labor is too onerous in a country whose people migrate by the hundreds of thousands in search of work. - 10 - 23. Apart from these aspects, providing mechanical equipment to Upper Volta makes the Government liable for continuing calls on foreign exchange, unless donors accept that they must provide spares throughout equipment life and that they must meet the eventual cost of new equipment as replacement. Without this commitment, and without a local ability to manage equipment- based operations effectively, plant is certain to be underutilized and to have a shortened working life. In turn, this will lead to high work costs, which often go unrecognized since equipment residual values are seldom cor- rected from original assumptions and, inevitably, economic reevaluation will give misleading results. The Road Impact Study 24. The study was planned to generate information which would have been interesting to Bank Group staff but the record shows that Government did not regard it as particularly important. The time and resources allocated to the study were, in the audit's opinion, too limited for the arrival at any sound conclusions. The Provisional Final Report (para. 24.1.6) prepared by the consultants states that "it is impossible to present more than a general pic- ture of the effects of road improvement- because there were too many extra- neous variables which masked the effects of road condition. The report itself has vanished into Government archives and when it was retrieved by the audit mission in December 1983 it was discovered that Government had still not commented on it. However, the Fourth Project plans to continue with a similar study. 25. The audit considers that under Voltaic conditions, the road selec- tion criteria set out in Part C of the January 1975 Issues Paper provided a sensible approach for the identification and inclusion of roads under- the project's work program. This approach could have been codified into a simple set of rules for initial road selection, with evaluation of a sample of the roads to be undertaken through the technical assistance to BPP. The effort to carry out a survey (SAR, Annex 4) which seems too complex for the level of statistical data likely to be available in the country would have been better used, as acknowledged in the PCR (para. 3.24), in making sure that simple traffic counts were carried out. The absence of traffic count data continues to hamper road planning even in early 1984, nine years after appraisal of the Rural Roads Project. Donor Coordination 26. Since the early 1970s complete lack of coordination among donors has been a characteristic of project planning and implementation in Upper Volta. Sporadic attempts to rationalize the volume and direction of aid flows have been made but they all failed because Government is not keen to encourage official exchanges among donor agencies and because donors are keen to preserve their independence of action. The audit mission discussed aid coordination with Government and donor agency officials and concluded that neither side has changed its long-held position. The audit believes that work executed in enclaves does not produce additive results and is in the long-term interest neither of the Government nor of the donors. - 11 - IV. CONCLUSIONS 27. During project inception and preparation, it was clear that since Upper Volta lacked human and financial resources for planning, improving and maintaining its rural roads it had to rely on external assistance. Donor agencies provided staff and mechanical equipment but rural road improvements were largely done on an enclave basis and with no reference to any overall plan. Maintenance was generally neglected and the condition of the rural road network remained unknown. No traffic counts were done on rural roads and there was uncertainty about the effect poor communications might have upon agricultural development. Up to the appraisal stage, project objectives were stated in clear terms: improve rural road sections that promise to have an effect on increased agricultural production; carry out routine and peri- odic maintenance on recently improved sections; help establish a service in the DPW so that, in the long run, the country will carry out with its own resources planning, upgrading and maintenance of all secondary roads. 28. At appraisal, project objectives began to lose their clarity. The Bank Group did not analyze the management and recurrent cost burden that the project, together with similar donor-assisted operations, would impose upon the Government. Despite Bank Group staff awareness of the poverty, emigra- tion, foreign exchange, staffing, and budgetary problems confronting the country, equipment-based techniques were selected for all work to be carried out under the project. The choice had three implications: first, the roads would have geometric characteristics suitable for the accommodation of the machines that would work on them but too generous for traffic volumes of 10- 20 vehicles per day; second, work would stop if one or more critical pieces of equipment were immobilized; third, equipment-based construction methods generate considerable demands on the foreign exchange reserves of a resource-poor country. None of the three implications seems to have been examined. In addition, the project's institution-building objectives assumed increased complexity: the methodology for road selection and evaluation became too intricate and a delicate balance was expected to be struck between the not-yet-established Studies Section of SERS and the not-yet-established BPP, while an Interministerial Committee would provide coordination. None of these expectations materialized. Furthermore, the project now provided for a complex Road Impact Study although there was no indication that the Govern- ment was interested in the exercise. 29. Increased complexity had a direct impact on implementation. Proj- ect start was delayed by 10 months because technical assistance staff was late in coming on board and because much time was spent repairing equipment taken over from other Bank group-assisted projects. Delays kept mounting up and the project was declared completed 30 months after the planned date with 61% of the planned improvements finished and with 90% of the planned mainte- nance. However, the end of the project saw SERS weakened rather than strengthened. - 12 - 30. Centralized administrations are acceptable in some societies but not in all and not often in sub-Saharan Africa. SERS is a typical example of a centralized road planning, construction and maintenance service and in the long-run its most important effect would have been the lowering of rural road construction standards. This must have been politically undesirable and rural road improvement and maintenance operations were decentralized in 19b0. The SERS head office is now part of the Ouagadougou Regional Directorate which has more office and workshop capacity than it needs. SERS branch operations in Bobo are integrated into the Bobo-Dioulasso Regional Directorate which continues not to have a clear responsibility for works carried out in its region. The Bank Group supported the creation of tie cen- tralized SERS and subsequently also supported its decentralization. TLe fact remains that even under the best circumstances decentralization entails some degree of duplication and requires a somewhat greater capacity than an efficiently run centralized operation. In Upper Volta, the resources were scarcely sufficient to support a centralized system, let alone a decentralized one. On the other hand, given the absence of any coordination of foreign aid and the time consuming procedures followed by central Government services, decentralization may be the way to get something5 done, albeit at the cost of duplication, efficiency and cost-effectiveness. 31. The basic reason why there is no coordination of foreign aid is that the Government does not have a clear idea of its own development strat- egy. In the case of transport, even a broad statement concerning targets which the Government has committed itself to pursue would have allowed donors to pick items and activities consistent with their respective assistance philosophies and crucial to the attainment of Government objectives. The Bank Group has been assisting the transport sector for almost fifteen years. The time has now come to alert the Government to the benefits of an indica- tive plan and to explore ways in which the Bank Group could assist the Government with plan preparation. It is however important that the useful- ness of such an exercise be fully understood by Government since planning initiatives promoted from outside very rarely bear fruit. 32. In the absence of a clear description of Government targets, donors will continue to operate in enclaves. This allows them to by-pass the in- ordinat-l!y long process which applies, for example, on procurement of goods and services through Government channels and which may take years for the purchase nf, say, spare parts. By operating in enclaves, donors adopt a ;ariattn of the decentralized approach and they do get a number of roads built. However, lack of resources in the recurrent budget may prevent these roads from being properly maintained. Thus, a new cycle begins when some aid nc-: ::.014 conceive a new project to deal with deferred maintenance, im- 5/ The Bank's Western Africa Regional Office feels that decentralization ma' have been premature in 1981 because Government finances ccticidentally declined and qualified staff were lost for political reasons. However, the new structure is now taking hold and in the event, decentralization was not an inappropriate solution. - 13 - provement and reconstruction while, in the meantime, poor road condition will impede road transport and constrain agricultural development. 33. The project set out to pursue fundamentally sound objectives and the audit investigated the extent to which the flow of benefits has survived project completion. In summary, the audit found that the SERS institutional capability has been compromised by the shift to decentralization; that the construction technology used was not, under Voltaic conditions, the most appropriate; that the project's recurrent cost implications are so large that it is doubtful whether Government can sustain the burden with its own resources. Recent socio-political developments have added new dimensions which deserve thorough study prior to the formulation of future Bank assis- tance to Upper Volta. -/Y-- -g /gek -15- UPPER VOLTA RURAL ROADS PROJECT (CR. 579-UV) PROJECT COMPLETION REPORT I. INTRODUCTION 1.01 Upper Volta is a sahelian country, surrounded by 1ali in the West and North, Niger in the East and Ivory Coast, Ghana, Togo and Benin in the South. It extends over some 274,000 km2 of which only about 90,000 km2 are potentially arable. With a total population of around 6 million, average arable land per person is less than 1.5 hectare, but still agriculture provides the livelihood of over 90% of the people. Close to one million young Voltaics seek a living abroad, most of them working as agricultural and menial laborers in Ivory Coast. Their remit- tances provide a considerable proportion of the small cash-incomes of their extended families. Even in 1980, GNP per capita for Upper Volta was estimated at no more than US$210 equivalent per annum. 1.02 In this landlocked country, the Government's first transport objective was to provide alternative access routes to the ocean via Ivory Coast, Ghana and Togo. After achieving this goal, it turned toward buil- ding an all-weather road network in order to further political integra- tion, provide improved access to agriculture markets, and link population centers. Since, with the exception of the trunk roads between the Ivori- an border and Bobo Dioulasso in the southwest, and Ouagadougou to Fada- N'Gourma and toward the Niger border in the east, traffic densities throughout the country are low, investments in roads must be considered carefully. Given the country's limited financial resources and institu- tional development, the burden of maintaining the network becomes an increasing problem as the road system expands. 1.03 The Association's operations in Upper Volta have been predomi- nantly in the agriculture sector, and the three transport projects to date have largely supported agricultural development. The purpose of the First Highway Project (Cotton Roads Project, Cr. 316-UV for US$2.8 mil- lion, in 1972, increased in 1973 by a Supplementary Credit of US$1.35 million), was to construct the Koundougou-Solenzo (70 km) and Hounde- Bereba (27 km) roads in the West Volta cotton area. The project was completed in 1975, except for the Hounde-Bereba road which was cancelled when it was found that cotton production in the region was insufficient to justify it. In the same year, the Rural Roads Project (Cr. 579-UV) was negotiated; it is the subject of the present report. - 16 - 1.04 The Third Highway Project (Cr. 653-UV, US$20 million, July 1976) provided for reconstruction to paved standard of the Banfora-Bobo Diou- lasso-Hounde road, vehicle weight control and traffic counting equipment, feasibility studies, and technical assistance. The project was completed in 1981 with construction objectives met fully and on time, but only modest success for the institution building aspects. Except for taxes, the project was fully foreign financed, since the Government obtained a further credit of US$3.5 million from the EEC and a US$2.1 million grant from OPEC to pay for local costs and cost increases. (PCR, June 15, 1982). 1.05 A considerable amount of rural road improvements were carried out as part of various IDA agricultural projects: the West-Volta Cotton Project (Credit 225-UV, US$6.2 million, 1971) included upgrading of 820 km of rural roads at an estimated cost of US$1.3 million; the Rural Development Fund (RDF) Project (Cr. 317-UV, US$2.2 million, 1972) and the Drought Relief Fund (DRF) project (Cr. 422-UV, US$2.0 million, 1973) foresaw road improvements on some 500 km that accounted for around 25% of the projects' funds; finally, the Bougouriba Agricultural Development Project (Cr. 496-UV, US$8.0 million, 1975) financed a road construction brigade that was to improve over 600 km of area roads and tracks at a cost of about US$1.6 million. Execution of the road components ranged from satisfactory (RDF/DRF results: 366/500 km constructed) to very good (West Volta Cotton: 270 km regravelled, 3,440 km graded; Bougouriba: 640 km regravelled, 1,490 km graded). The equipment financed under these projects was taken over by Public Works to be used on rural roads mainte- nance under Cr. 579-UV (see para. 3.02 below). 1.06 The Fourth Highway Project,(Cr. 1164-UV, US$46 million) negotia- ted in 1981, has been declared effective in August 1982 only. It pro- vides for periodic maintenance and the incremental cost of routine main- tenance of the priority classified roadnet; for the upgrading of some 1,200 km of rural roads serving integrated agricultural development areas; for technical assistance; and for road impact monitoring and pilot community-based maintenance schemes. 1.07 The present report is based on Government progress reports, Bank supervision reports and information collected during a project super- vision/completion mission in March 1982. - 17 - II. PROJECT PREPARATION AND APPRAISAL A. Preparation and Processing 2.01 The Rural Roads Project was prepared by a French consulting firm financed by the Fonds d'Aide et de Cooperation (FAC, France'. The con- sultant's report was issued in May 1973. It proposed the creation, in the Directorate of Public Works (DPW), of a Special Service that would execute the proposed project and, as its capacity developed, gradually take over the responsibility for the programming and execution of all rural road construction and maintenance activities in Upper Volta. An indicative three-year road construction and maintenance program was de- fined for the new Service which was to be known as the Secondary Road Maintenance Service (Service d'Entretien des Routes Secondaires, SERS) in parallel to the already existing National Road Maintenance Service (Ser- vice d'Entretien des Routes Nationales, SERN). 2.02 The project was appraised in December 1974. Negotiations were held in Ouagadougou five months later (May 21-24, 1975) and the appraisal report was issued on June 6, 1975. The project was presented to the Board on June 16, 1975. Signature of the Credit Agreement for US$7.5 million took place on August 13, 1975 and the Credit became effective as of February 11, 1976. B. Project Description 2.03 The Rural Roads Project as appraised had two main objectives: (a) to support agricultural development through the improvement of some 1,200 km of rural roads and maintenance of an addi- tional 2,100 km of existing feeder roads constructed ear- lier under various IDA-financed agriculture projects; and (b) to establish and strengthen the institutional framework within DPW for the rational expansion and continuing main- tenance of the country's rural road network. A third component foresaw an evaluation of the impact and effectiveness of the road components included in the agricultural and rural development projects financed by the Association over the previous five years. 2.04 The proposed road improvements included: (i) reshaping of roadway; (ii) regravelling of selected sections; and (iii) building or reinforcing of drainage and crossing structures. - 18 - Two levels of improvements were planned as justified by present and ex- pected future traffic levels (Table 1). Type A roads were estimated to be justified at traffic levels of more than 15 vpd; type B roads should have at least an average daily 6-8 vehicles. Only some 250 km (134 km Type A; 116 km Type B) were firmly identified at appraisal; they were to constitute SERS' first year construction program. An additional 1,280 km had been examined by consultants and discussed with the Government for upgrading during subsequent years. The road network to be maintained was specified in detail in the appraisal report: it initially consisted of 1,210 km of classified Regional Roads and 870 km of tracks improved under IDA agriculture projects, to which project-improved roads were to be added after upgrading was carried out. 2.05 The institution designated to take over rural roads responsibi- lity, SERS (cf para 2.01), was set up within DPW in 1975. It was to be strengthened through the addition to its personnel of three technical assistants (two road engineers and one administrator/accountant for equipment management); and was to be provided with construction equip- ment and vehicles (partly to be purchased under the project, partly to be taken over from IDA agriculture projects) to form two construction and two maintenance brigades with which to carry out the first three years' work program. 2.06 At the same time, the Bureau of Planning and Programming (Bureau de Planification et Programmation, BPP) within DPW would cooperate with SERS in the development of consistent technical and economic criteria for establishing definitive annual road construction programs. These might differ from the indicative three-year appraisal program if changing deve- lopment priorities of the regional development authorities (Organismes Regionaux de Developpement, ORD) necessitated changes in the initially established program. 2.07 To ensure responsiveness of the road improvement program to agricultural and social development priorities, an Interministerial Tech- nical Committee (ITC) comprising representatives of the various minis- tries concerned (Rural Development, Health, Education etc.) and of the ORDs was to be established. This committee would meet annually to review and approve the program prepared by BPP, in accordance with technical and economic criteria acceptable to the Association. 2.08 Small contractors were to be employed for the construction of minor drainage and crossing structures, and for haulage of road surfacing materials. This was to be done in cooperation with the Office for the Promotion of Voltaic Enterprises (Office de Promotion de 1'Entreprise Voltaique, OPEV), a Government agency created in 1971 and financed by several donor countries, to stimulate the development of local induq- tries. - 19 - 2.09 The road impact study was to provide empirical measures of the impact of road improvements in the main agricultural zones and socio- cultural environments in the country. In order to learn from earlier project experience and so as to facilitate data collection on agricul- tural production parameters, sample roads were to be chosen among those improved as part of various IDA financed Agricultural Projects especially in the southwest of the country, for which a considerable amount of data was available. The study was to evaluate road project preparation, in- plementation and supervision under earlier projects; establish the costs of different construction technologies (labor/equipment intensive; force account/contractor); and compare estimated with actual costs of the road improvements. It was also to determine the level of road user savings ascribable to the improvements and establish their distribution between transporters, traders and producers. The findings of the study were to be synthesized into an improved model for rural road evaluation to be used in future years' construction planning and programming by SERS. C. Cost Estimates and Financing 2.10 The project was estimated to cost US$8.5 million equivalent net of taxes with a foreign exchange component of US$6.6 million equivalent or 78% of total project cost. In addition to local costs of US$1.9 mil- lion equivalent, taxes and duties would amount to US$1.7 million. The credit of US$7.5 million financed all foreign costs, as well as US$0.9 million equivalent of local costs. The remainder was to be provided by the Government. 2.11 Credit funds were to cover the following goods and services required to carry out the proposed construction and maintenance program: (a) 100% of the foreign exchange cost of equipment, and of the necessary spare parts materials and supplies, for two re- gravelling brigades and two culvert-building units; (b) 100% of foreign exchange costs for technical assistance and consultants; (c) 64% of the cost of civil works by contractors; and (d) 50% of operating expenses for betterment and maintenance work under the project. 2.12 At appraisal, it was foreseen that project execution would begin in mid-1976, and the road improvements were expected to be completed in three years. The original closing date was December 1979; this was sub- sequently extended to December 1981. - 20 - D. Covenants and Sideletters 2.13 The Credit Agreement and sideletters specified the following actions to be taken by Government: (a) adequate maintenance of the road network including =ainte- nance and renewal as required of road maintenance equipment (Section 4.02 of the Credit Agreement); (b) classification of the rural roads to be constructed or maintained by the project (Section 4.03 of the Credit Agreement); (c) allocation of at least CFAF 200 million for the maintenance of the rural road network in the annual national budget of the year following completion of the project, and such increased amounts as may be required to ensure continued satisfactory maintenance during succeeding years (Section 4.04 of the Credit Agreement); Cd) selection of roads to be included in second and subsequent years' construction programs from among those with an esti- mated economic rate of return of at least 10%, calculated according to criteria and following a methodology accept- able to the Association (Side letter of August 13, 1975, Annex A); (e) inclusion, in the proposed annual road upgrading program of a description of the extent to which SERS intended to use local contractors and self-help community action in carry- ing out the program (Side letter of August 13, 1975); and (f) investigation of the scope for part-financing of rural road network maintenance by the local communities served by the roads (Side letter of August 13, 1975). III. PROJECT IMPLEMENTATION A. Project Start-up 3.01 The various institutions that were to be in charge of project execution were all formally established well before the date of credit effectiveness. The ITC, which was modelled on a similar committee set up for the DRF project, was established in March 1974. SERS, the entity that was to carry out project works, was created in January 1975 and was given temporary offices within MPW's Equipment Pool facilities located in - 21 - an industrial zone on the outskirts of Ouagadougou. The BPP finally, was established in June 1976, following the signature, in April of that year, of an agreement between Government and the UNDP, who were financing the technical assistance required to establish the Bureau. 3.02 Even so, actual start-up of works did not occur by mid-76 as estimated, but was delayed well into Spring 1977. This was due to slow acquisition of equipment by SERS and to the poor condition of the equip- ment it received from other brigades. The ex-RDF-Project brigade was the first to be transferrred to the new Service in July 1976, but it required such extensive equip-ant repairs that it was not operative before January 1977. The ex-West Volta Cotton Project brigade was transferred to SERS only in early 1977 and had to undergo important repairs as well. As for the new equipment ordered directly for SERS under the project, delivery was delayed to spring 1977, due mainly to cumbersome local bureaucratic requirements and insufficient familiarity with Bank procedures. Although bidding documents for the equipment of the two new regravelling brigades were ready at appraisal and approved by the Bank on January 8, 1976, there was almost a full year delay between bid opening (April 2, 1976) and equipment delivery (February-March 1977). 3.03 Part of the explanation for the delay in delivery of equipment, spares and supplies was the delay in budgetary provisions for the newly created SERS; only in September 1976 did Government establish a special account for SERS with an allocation of CFAF 400 million, from which pay- nents for the more important initial expenditures of the Rural Roads Project could be met. Even then the administrative circuit between or- ders and final payment to suppliers remained a lengthy one. 3.04 Recruitment of local headquarters personnel for SERS was almost complete by the time equipment first became operational in April 1977. Recruitment of local field personnel was much more difficult: a lot of construction activity was going on in the country, and wages paid to experienced operators by private contractors were considerably higher than Government rates. Between April and November 1977, SERS lost 27 skilled operators to the private sector, and one of the four regravelling brigades was immobilized for lack of field personnel. 3.05 The technical assistance road expert (deputy to the head of SERS), was provided under French aid to MPW and the project was assigned an experienced local accountant, leaving one technical assistant position to be filled under the project. However, the incumbent deputy head of SERS at project start-up soon left and a lengthy period elapsed before the appointment of a successor. The appointment of the IDA financed expert who was to act as the Chief of SERS's Study Section was also delayed until early in January 1978, which was unfortunate as his pre- sence in the project was crucial to start proper programming and tech- nical preparation of all project road improvements. - 22 - 3.06 The contract for carrying out the road impact study was awarded to a UK consulting firm, in association with a Voltaic firm who provided some technical experts and all field personnel. Fieldwork for the study was launched in January 1978. B. Project Execution (a) Execution of Road Improvements by Force Account Brigades 3.07 Road improvement works started towards the end of April 1977. The ex-West Volta Cotton Brigade started work on the Samorgonan-Kourouma Section of the Orodara-Kourouma-Ndorola Road (101 km) included in the first year appraisal program; this road was important for the IDA financed livestock Develoment Project (Cr. 557-UV) in that region. The ex-FDR Brigade started with intensive maintenance ("reprofilage lourd") operations on the Banfora-Sindou Section of the Banfora-Baguera Road: construction of the 33 km Sindou-Baguera stretch (again part of the SERS first year program) first necessitated maintenance on the 52 km Banfora- Sindou section improved with FED financing some two years earlier. Re- construction of the remaining two "first year SERS program" roads (Lato- din-Kaya, 69 km and Boulsa-Zorgo, 48 km) was taken over by Canadian Aid (CIDA) and US-AID respectively. So the two newly equipped SERS brigades started their operations on two "indicative program" roads in the east of the country: Fada N'Gourma-Comin Yanga (50 km) and Moaga-Bissiga (24 km) near Koupela. 3.08 Of the programmed 1977 construction output of 250 km, 104 km were actually realized between April and December. Given the poor condi- tion of the inherited equipment for two of the brigades and the repeated loss of personnel to the private sector, this was quite a creditable performance for the newly established SERS. 3.09 Construction outputs increased over the next two years as bri- gade personnel stabilized and replacements arrived for some of the worn- out equipment. In all, 165 km were improved during 1978 and 242 km in 1979. During 1980 however, construction output declined again to 140 km and, more worrisome, to under two-thirds of this (85 km) during 1981, when SERS started to be affected by the decentralization of DPW proposed during preparation of the Fourth Highway Project. Thus, a total of 736 km were improved under the project, compared to the indicative program of 1,200 km expected at appraisal (for details, see Table 2). 3.10 Some of the shortfall in "finished" kilometrage is more apparent than real. While two different construction standards had been foreseen at appraisal, both featuring spot gravelling only, the roads actually constructed were fully regravelled to an average thickness of about 20 cn due in part to the considerable volume of earthworks associated with minor structures. The main reason for continuous gravelling was to avoid -23 - the need for scarce and costly managerial inputs that adequate supervi- sion of appropriate use of variable standards in the field would have required. In addition, road surfacing materials were readily available on all work-sites and the possible savings to spot gravelling were largely offset by the generally better rideability of a uniformly gravel- led road, the reduction of road closures during the rainy season, and lower maintenance requirements. (b) Construction of Bridges and Small Structures 3.11 The project was intended to contribute to the development of small contractors for civil works, mainly in the construction of small drainage works and materials supply and haulage. However, OPEV's perfor- mance was disappointing. By the end of 1978, it had not managed to pro- pose any small or medium-sized local enterprises interested in and/or capable of carrying out this work. With the lag between road works and drainage works steadily increasing, SERS itself launched two rounds of local competitive bidding in early 1979. The results were disappointing: only large foreign and domestic enterprises submitted proposals, all of them at prices of about triple the cost estimated at appraisal. 3.12 SERS who had used the equipment initially foreseen for two small culverting units to supplement the brigade equipment received from the agricultural projects, then established a "structures brigade" and started carrying out the drainage works by force account in 1979. Gradu- ally, the backlog of structures on earlier roads was caught up on, while on all new worksites structures were programmed to be completed just ahead of earthworks and regravelling. At the end of the project period (December 1981), drainage works were completed on 90% of the total.736 km improved under the project. (c) Execution of Road Maintenance by Force Account 3.13 Maintenance of feeder roads improved under earlier IDA-financed and other projects was an important part of the project. The appraisal report specified some 800 km of Type A roads and 1,280 km of type B roads that should be maintained by SERS. The former included the 70 km Koun- dougou-Solenzo road constructed under the First Highway Project, 100 km improved under the West Volta Cotton Project, 470 km upgraded with FAC or FED financing, and about 170 km in the Bougouriba Project Area. Type B roads included 400 km of classified Regional Roads (of which 100 km in the Bougouriba area) and around 880 km of non-classified tracks. Of these, 350 km had been improved under the West Volta Cotton Project, 190 km by the RDF/DRF Projects, and 340 km by the Bougouriba Project. - 24 - 3.14 With four maintenance teams, SERS carried out maintenance on all of these roads except for the 600 km in the Bougouriba region, since they could be more efficiently maintained by the existing Bougouriba Project brigade. At the conclusion of the first Bougouriba Project in late 1980 the remaining equipment and the maintenance responsibility were trans- ferred to SERS who, after considerable equipment repair and renewal, is now executing road maintenance in Bougouriba Project area as well. 3.15 Details on the road network maintained between 1977-1981 and the frequency and type of maintenance carried out are in Table 3. In all, intermittent maintenance operations covered some 3,300 km of roads and tracks. Of these 1,300 km, or just under 40% received heavy maintenance (several grader passes, partial regravelling and clearing of drainage ditches and culverts). The remainder received only light maintenance, essentially grading and clearing of ditches. Eight of the roads constructed under the project are reported among the roads maintained by the maintenance brigades; others were maintained by the construction and/or structures brigades that were sent back to roads where gravelling had been completed in the first two years, but where the structures were put in place in the later years of project execution (cf. para 3.12 above). 3.16 Maintenance outputs during the five year project period show a pattern parallel to that for the construction work: 380 km in 1977; 710 km in 1978; 1,510 km in 1979; 2,070 km in 1980; and a decline to 850 km in 1981. Start-up of activities was slow; outputs approached apprai- sal estimates only in the third and fourth year, and dropped heavily thereafter. (d) Annual Construction Programming 3.17 Of the 250 km of project roads that had been identified for improvement during the first year of SERS operations, 117 km were subse- quently included in other programs. The ITC was therefore convened in November 1976 to examine and approve the inclusion of six other roads (about 320 km) in the 1977/78 construction program. One of these was a road included in the indicative program for later years; two were roads that had been summarily improved with RDF Funds. The remainder were requested by the ORDs (see Table 2 for details). The Association approved inclusion of the first three roads (about 140 km) and requested that the very summary evaluation procedure to justify reconstruction be improved. The method applied essentially provided no more than an esti- mate of existing (1975) tonnage of marketable agricultural production (without road improvement) and expected (1980) marketed volume 1/in each road influence area. SERS and BPP were asked to provide an estimate of the construction cost for each proposed road section based on detailed l/ The appraisal report had estimated that annual traffic volumes above 2,500 tonnes justified upgrading of roads to Type A standard. - 25 - engineering surveys by its Studies Section, and a rate of return calcu- lation that reflected not only the cost of road construction and main- tenance, but also the cost of incremental modern agricultural inputs necessary for increasing crop yields in the road influence areas. 3.18 From mid-1978 to mid-1981, the SERS "Studies Section" carried out detailed topographical surveys for nine roads 1/ and established cost estimates for each road section which reflected the cost of minor re- alignments and of the varying types and numbers of structures needed. The BPP, where two UNDP-financed technical assistants worked with two Voltaic counterpart economists on a range of subjects, developed a simplified producer surplus model 2/ to calculate a rate of return for each project road based on agricultural production estimates provided by ORDs. To ensure an uninterruped work program for SERS, road evaluations were submitted to the Association on a case-by-case basis, rather than as annual program packages, since it proved extremely difficult to convene the ITC in time for approving these. SERS in fact worked directly with each of the ORDs concerned in establishing construction priorities. The procedure worked reasonably satisfactorily, but BPP did little to system- atize and control the ad-hoc population and agricultural production esti- mates for the different road influence areas provided by the ORDs, on which evaluations were based. Estimated Cost/km ERR 1/ Road Sections Evaluated Approx. km (Mio. CFA) % Tougan-Nohoro 85 2.01 12 Fara-Ouessa 70 2.95 13 Batie-Kupere 60 2.70 12 Banfora-Mangodara 100 2.52 17 Kombissiri-Gana-Toesse 35 1.92 14 Ouarkoye-Fakena 8 2.35 18 Toumousseni-Loumania 85 2.95 25 Sindou-Tourny 34 2.65 16 Sindou-Banzou 57 4.04 21 524 2/ Based on Bank Staff Working Paper 241, "The Economic Analysis of Rural Roads." - 26 - e) Road Impact Study 3.19 At DPW insistance, geographical survey coverage of the road impact study was extended to include roads in all regions of the country. The study budget of US$257,000 (equivalent to 40 man-months of consultants' and 80 man-months of interviewer time) limited the sample of road sections studied to fourteen; within their influence areas, samples of 20 markets, 80 villages and 320 farm-households (for details see Table 4) were to be surveyed over a three-year period to monitor impact of road improvements on transport availability and costs; to measure the extent to which transport cost savings were passed on to agricultural producers, and to establish whether improved market access stimulated agricultural production and improved farmers' income. Survey intensity was low: one road condition survey per section; two rounds of household/village/mar- ket surveys; and one or two rounds of traffic counts (one week's dura- tion) over three years. Survey implementation and study management were competent and timely: questionnaires were carefully prepared and field tested; interviewers were locally recruited and trained by the Voltaic consultant; and survey work was carried out to schedule in 1978 and 1979. The planned 1980 round of surveys was deferred when preliminary comparative analysis of the first two years' results came up with only very few statistically meaningful measures of change in relevant para- meters. Instead, consultants concentrated on establishing an ORD level data base (population, agricultural production, social services) from existing sources, that could be updated regularly and used consistently for the "ex-ante" road evaluation exercises being carried out by BPP (cf para 3.18 above). 3.20 The study produced five reports and two Technical Notes in all 1/. The technical surveys addressed three main concerns: (i.) the availability of suitable materials for rural road construction amd main- tenance; (ii) the adequacy of adopted road improvement standards for the traffic the Upper Volta rural roads actually carry; and (iii) the appro- priateness of the road construction techniques used in the Upper Volta socio-economic context. Concerning (i), the study concluded that in all regions in Upper Volta lateritique gravels with a CBR of 20-30, and usu- ally above 30, can easily be found. However, suitable materials had not always been used on all road sections reconstructed, and simple 1/ (a) Inception Report February 1978 (b) First Technical Report April 1978 (c) Second Technical Report May 1979 (d) Draft Economic Report September 1979 (e) Draft Final Report September 1980 - 27 - laboratory test training for technicians was proposed to improve material selection on project sites. Concerning (ii), the study concluded that, in view of the nature of surfacing materials used and prevailing climatic conditions, it was desirable to have wearing courses constructed to a thickness of 20 cm: this supports actual SERS practice against earlier Association advice to reduce gravel thickness on low trafficked sec- tions. Concerning (iii) the study concluded that labor intensive road construction, while technically and economically feasible, would necessi- tate such major changes in SERS' organization of works and composition of workforce, that it was unlikely to be introduced on a large scale. Labor intensive methods could however, be envisaged for particular well-defined tasks, such as the building of small structures. Similarly, it would be not only technically and economically feasible to carry out a significant part of road maintenance labor-intensively, but also politically desir- able to do so. This was because (a) DPW would not, in the foreseeable future, dispose of adequate means to maintain its 8,700 km of classified roads using equipment intensive methods; (b) maintenance work in con- trast to construction offered long-term employment for a steady number of workers, and much of it could be done in agricultural "off-peak" per- iods; (c) maintenance by local labor would minimize foreign exchange expenditures and increase incomes in the road influence areas; and (d) there was a possibility of mobilizing local contributions to the upkeep of the newly constructed roads in the form of partly voluntary labor: in almost half of the 80 villages included in the socio-economic surveys, people had indicated their willingness to make such contributions. A pilot-project was proposed to test the practicality of introducing commu- nity based maintenance on a larger scale; it has not been implemented so far due to DPW's benevolent neglect of the proposition. 3.21 The socio-economic surveys attempted to monitor three main as- pects of hypothesized road impact: (i) changes in level and types of traffic, and in transport charges following road improvement; (ii) changes in agricultural production patterns, production technologies and in volume of products marketed associated with road improvements; and (iii) changes in availability and use of services (market, extension, education, health, etc.) in rural areas following road improvements. Concerning (i), the two rounds of traffic surveys in 1978 and 1979 esta- blished traffic levels at two points in time.1/ But coverage was too scanty for any meaningful trend to be established in year to year traffic development or between project road and control roads, especially since 1/ Fourwheel vehicle traffic observed ranged from 2-46 vpd (mean: 10.5 vpd); two wheel vehicle frequency was between 2-300 vpd, (mean: 115 vpd). - 28 - the road improvements under the agriculture projects had been very modest and had sometimes been carried out several years earlier. However, the study did not even attempt to relate traffic levels to actual road condition. Transport charges for passengers and freight on the roads surveyed followed no discernible pattern related to either road condition or distance travelled. Variations in passenger transport charges ranged from CFA 7.7-12.5 per km, and those for 100 kg of freight from CFA 4.0- 12.5 per kn, and were apparently being set in terms of what the traffic would bear. 3.22 Concerning the agricultural production impact of the roads ((ii) above), the village and farm household surveys showed no significant differences in volumes produced and marketed between study and control zones. The year to year differences in volumes of agricultural produc- tion reported for the localities surveyed were so improbable that the initially planned third-year round of surveys was cancelled. To measure changes in agricultural production reliably, the study would have needed to include a large sample of systematic yield surveys over at least a five year horizon, and even then, the large production differentials due to climatic factors may have outweighed production changes due to the road. 3.23 Concerning availability of, and access to, services in study zones and control areas, the surveys established some significant differ- ences. The level of marketing was somewhat higher in the influence areas of improved roads: 40% of farmers reported having sold to traders vs 21% in control areas. Visits by extension agents were reported by 60% of farmers in study areas vs 47% for the control zones. But the latter had a higher rate of visits by veterinary personnel (33%) than farmers in study areas (23%). Income from non-agricultural sources was considerably more important in the study areas than in the controls: the number of persons employed outside agriculture in the former increased by 31% bet- ween 1978 and 1979, compared to only 2% in control areas. 3.24 In summary and with hindsight it can be concluded that the approach used to derive measures of incremental agricultural production due to the road was unlikely to yield reliable results. It would pro- bably have been preferable to concentrate on repeated origin/destination surveys and on intensive monitoring of transport availability and cost in relation to the seasonal state of maintenance of the roads. Supplemen- tary village and household surveys should have concentrated on indicators of socio-economic development that are readily verifiable by the inter- viewers. Instead of endeavouring to establish production estimates from a small sample of farmer interviews, it is preferable to base producer surplus based economic return calculations on available agricultural statistics. While these are perhaps not as disaggregated as would be - 29 - desirable, they can nevertheless be regularly updated and, hopefully, improved in reliability over time. BPP should, under the Fourth Highway Project, set up a program of origin/destination surveys on a represen- tative sample of rural roads, and update the regional-level population and agricultural production estimates, provided by the study for the 1977-78 crop year, to serve as an acceptable basis for "ex-ante" road evaluations for annual construction programming. C. Implementation Schedule 3.25 At appraisal, the project was expected to start in July 1976 and be completed after three years at the end of June 1979. Effective pro- ject start-up was delayed by 10-months, and slow project implementation during the first year of operation further increased the lag. The pro- ject was completed with a thirty month delay at the end of December 1981. The closing date was deferred twice (to June 30, 1981, and December 31, 1981). D. Project Costs 3.26 Appraisal estimates and actual project costs are summarized in Table 5. Actual costs are less than two percent above the appraisal estimate inclusive of contingencies, since works were stopped when funds were exhausted; 61% of the kilometrage to be constructed was implement- ed, but to a higher standard than foreseen at appraisal, and close to 90% of the maintenance was carried out. Actual share of local costs in total project costs was slightly higher than estimated at appraisal (25% vs 22%). IDA financed 52% of these vs. 47% estimated: the increase is due to the significant local cost share for the impact studies, which em- ployed two local organizations alongside foreign consultants. 3.27 Costs of equipment were some 11% higher than estimated: more replacements were needed for the ex-agriculture project brigades trans- ferred to SERS than had been foreseen. Expenditures for spare parts and fuel exceeded appraisal estimates by one percent only. No works were contracted out, and the related allocation mostly served to finance the force account structures brigade of SERS. SERS personnel costs were double those foreseen at appraisal. They reflect the long (five-and-a- half compared to three years estimated) implementation period, and the cost of executing drainage works with project personnel. Head office costs at appraisal were expected to represent some 16% of total local staff costs; actually, head office (including central workshop) person- nel accounted for at least 25% of total staff expenditures. - 30 - 3.28 Technical assistance costs under the project were half those estimated at appraisal: they covered three years service of one road engineer/chief of Studies Section rather than two road engineers and one administrator/accountant as foreseen at appraisal. However the Proi'ct benefitted from five years' service of a road engineer/trainer f4-.arced under the West Volta Agricultural Development Project (Cr. 70-UV) and stationed with SERS in Bobo-Dioulasso. A certain proportion of the cost of the UNDP-financed technical assistance to BPP should also be added to this cost component, but it is not possible to give a realistic estimate of the share of their time spent working for SERS. The same is true for the FAC-financed deputy-head of SERS, who devoted a considerable part of his time to the project. The true cost of technical assistance may in fact be almost double the appraisal estimate. 3.29 The budget for the road impact study was underestimated at appraisal, and actual costs amounted to 175% of those estimated. With hindsight, it would have been desirable to allocate additional funds in 1981 for traffic counts on all project roads, since BPP have so far been unable to organize any systematic traffic surveys either on the national, or on the rural roads. 3.30 Analytical cost accounting data for the different construction and maintenance activities at the site level were collected starting in 1978 and provided reasonably complete information for 1979/80. Record keeping and report preparation became sporadic after project management changed and technical assistance had left. Based on the partial informa- tion available, an attempt has been made to estimate the cost of mainte- nance operations and, after deducting these and the costs of the study from total project costs, to derive an average per kilometer cost for the roads constructed. It would seem that the average per kilometer expendi- ture (current prices) for road maintenance was CFAF 35,000 or just under US$150. The cost per kilometer of road improved can then be estimated at about US$9,680 equivalent net of taxes, roughly 60% higher than the 1975 estimate including contingencies for the average per kilometer construc- tion cost (cf. Table 6 for details). Taxes add another 20% to this estimate. E. Disbursements 3.31 Expected and actual disbursements per quarter and per Credit category are shown in tables 7 and 8, respectively. The last disburse- ment was made in September 1982. - 31 - F. Technical Assistants' and Consultants' Performance 3.32 Technical assistants to the SERS Studies Section and to the Bobo-Dioulasso SERS Regional Office performed very satisfactorily during their assignments. On-the-job training of local counterparts was among the most important aspects of their assignments and was carried out quite successfully. 3.33 For the technical assistants to BPP, who were to carry out the socio-economic evaluation of road sections newly proposed for upgrading by SERS, this was only one of many other tasks assigned to them. They developed a useful mini-computer producer surplus model for Upper Volta, and they trained the counterpart economist (s) in its application. However, they did not themselves establish an ORD-level data base, that would have allowed some control of the agricultural production estimates proffered by_ORD-agents for particular road influence areas; nor did they cooperate at all closely with the consultants in charge of the Impact Study when the latter engaged in this work. 3.34 Consultants in charge of the Road Impact Study tackled a diffi- cult task imaginatively, but perhaps over-ambitiously. The Voltaic "jun- ior partner" firm resisted being restricted to interviewer training and supervision and after some initial difficulties were associated with important aspects of survey design, analysis, and report preparation. The consultants carried out surveys competently and presented reports on schedule. Each study phase was discussed with DPW as well as with the Association, and changes in orientation based on field experience were jointly agreed upon. But the final report, having been published at the time of the military takeover of government, has never been discussed with the Borrower or the Association. G. Borrower's Performance 3.35 The Borrower's performance continued satisfactorily for the first three years of project implementation. Daring this period execu- tion of works steadily improved, project financial accounting was esta- blished, cost accounting was introduced and disbursement requests were presented reasonably regularly to the Association. Overall SERS' output presented a creditable performance for a newly created service in a country beset with as many institutional and socio-political difficulties as Upper Volta experienced during the period of project execution. 3.36 Two factors can be identified as causing the decline in cons- truction and maintenance outputs (cf. paras 3.09 and 3.16 above) during the last eighteen months of project execution. First, SERS was disrupted by the decentralization of DPW carried out under preparation of the - 32 - Fourth Highway Project: this led to the rotation of all the counterparts trained by the project from positions where they had been performing competently into new posts where their responsibilities still have to be defined. Since technical assistants were leaving as scheduled at about the sane time, some jobs were simply without incumbents and project per- formance began to slide backwards. At the end of 1980, SERS was further weakened by a change in management following the military take-over in the country. This prompted several months of investigations into project accounts during which project activities were practically suspended. The head of SERS was promoted away and no longer dealt with its day-to-day affairs, but administrative authority to engage project expenditures was not officially transferred to any successor until well into 1982. Al- though more funds than had been estimated were available in the credit due to the significant increase in the dollar exchange rate, only a trickle of disbursement requests were presented to the Association. Suppliers in Upper Volta remained unpaid and stopped deliveries of spare parts and fuel; salaries were paid late, personnel attendance became erratic and construction outputs dropped sharply. Although a number of supervision missions repeatedly outlined the measures needed to redress the situation, DPW response has been disappointing and slow. H. Compliance with Covenants 3.37 The Government, through the operation of SERS made a reasonable, if not completely satisfactory effort to maintain the rural roads improv- ed under this and a number of earlier agriculture projects. However, official classification (cf. para 2.12 above) of all improved roads, the administrative pre-requisite for including roads into the net under DPW's maintenance responsibility, has not yet been carried out. The allocation of CFAF 200 million for rural road maintenance in the annual budget of the year following project completion has not been made specifica3ly, but the global road maintenance budget allocation is in line with esti"ated needs. 3.38 Roads proposed for inclusion in the yearly construction programs have been selected in accordance with criteria acceptable to the A-socia- tion. A study was commissioned by SERS to establish a practical approach to involve local communities in road maintenance. No follow-up pilot works have been organized so far; these should be implemented under the Fourth Highway Project. - 33 - IV. INSTITUTIONAL DEVELOPMENT 4.01 Among the main objectives of the project was the creation, in- side DPW, of a capacity to evaluate, progran and execute rural road cons- truction and maintenance. The preceding section on project execution has traced SERS' development as an "institution" over the period of project implementation. Acquisition of equipment and training of personnel for making the brigades operational took considerably longer than anticipa- ted. However, as project administration learned to function more effi- ciently despite of cumbersome administrative requirements and succeeded in training and retaining reasonably competent brigade personnel, initi- ally poor construction outputs gradually approached appraisal targets. For two years (1979/1980) the project functioned satisfactorily with annual production outputs averaging 200 km of road improvements and 1,800 km of maintenance. 4.02 The consolidation of SERS as an institution was visibly re- flected when it moved in mid-1980, to its own office facilities in a new building provided by the Ministry of Public Works. This cost a total of CFAF 260 million to construct aad an additional CFAF 24 million to fur- nish. Besides SERS, it houses the Directorate of New Works, and the administration of various foreign-financed rural road projects (cf. para 4.06). In 1981 the new offices were complemented by extensive workshop facilities on the same site, where all SERS' heavy repairs now can be handled. The workshop cost CFAF 48 million to construct. The project contributed none of the financing for SERS headquarters. 4.03 But just as the physical infrastructure for a centralized SERS became fully operational, DPW institutional structures began to be rear- ranged into a more decentralized pattern. SERS head-office became part of the Ouagadougou Regional Directorate which now has more office and workshop capacity than it needs. SERS branch-operation in Bobo is being integrated into the Bobo-Dioulasso Regional Directorate, which is build- ing up office staff to reflect its new importance, but still has no clear responsibility for works carried out in its region. SERS construction and maintenance outputs declined sharply in 1981. It is now clear that due to the combined effects of political change in the country, reorganization in the Public Works sector and delays in the date of effectlveness of the follow-on Fourth Highway Project caused inter alia by non-availability of IDA funds, more than two years' construction output will be lost, before the regional directorates who are henceforth responsible for all rural road construction and maintenance in their areas, will begin to operate efficently. - 34 - 4.04 Construction and maintenance programming for 1982, beyond roads to be constructed in support of IDA Agricultural Development Projects, is embryonic at best. Since the reorganization of DPW, it has been diffi- cult to interest DPW personnel in the results of the impact study which provides an improved, if by now somewhat outdated agricultural data base for road evaluations that should be continuously updated by BPP for future use. BPP's efforts to establish a country wide system of traffic counts also have so far remained completely ineffective: of 50 automatic traffic counters ordered for use on the trunk network in 1976/77 and delivered in 1979, a total of eight have so far been installed. Half of them have been destroyed or stolen, and none has produced any exploitable results. The only traffic counts so far available are from 14-day manual counts at the four counter locations, undertaken to derive a sample dis- tribution of vehicle types using the main roads. No counts at all have been carried out on secondary or regional roads. 4.05 Under the present circumstances, the planned integration of all non IDA-financed rural road construction activities under SERS has become a somewhat moot point. The desired coordination of rural road works by the many international, regional and bilateral organizations involved in the sector is likely to continue in an ad-hoc fashion. The various donor organizations execute road works within fairly well-defined regions, but according to different criteria and to differing standards. The United Nations Sahelian Office (UNSO) finances and controls projects identified through the CILSS (Comite Permanent Inter-Etats de Lutte contre la Seche- resse dans le Sahel). It has provided operating funds for a road cons- truction brigade provided by the Government and staffed by Upper Volta Army Engineers, which started work in 1978 on the 107 km Ouahigouya-Djibo road. By the end of 1981, about 90 km have been completed. The Canadian Aid (CIDA) program finances improvement and maintenance of some 380 km of roads. It provides equipment, 25 Canadian plant operators and a large part of operating costs for the brigade. Construction started in 1979 on the 250 km Barsologho-Kaya-Latodin-Tougan road which was completed in 1980. Works are now going on south of Ouagadougou. United States Aid (US-AID) has been particularly active in the Eastern region. It financed equipment for a brigade plus technical assistance for its operation starting in September 1979. At the end of 1981, this brigade has improved a total of 164 km (Namounou-Lagobow, 40 km; Comin Yanga-Diabo, 48 ka; Fada N'Gourma-Bilanga, 76 km). The French Fonds d'Aide de Cooperation (FAC) has been providing road construction equipment to DPW for years. It financed a special road brigade for SERS in 1978 to construct the Kongoussi-Djibo road which has been maintained under the present project. FAC also finances a total of 27 technical assistance post throughout the Ministry of Public Works which include the Deputy Chief of SERS. - 35 - 4.06 In the early years of project execution, effective control and coordination of all of these activities was clearly beyond SERS' administrative capacity. The UNSO, CIDA and US-AID brigades thus started as self-sufficient "enclave" operations. They are likely to continue in this way, since it allows them to by-pass administration red tape concerning procurement/employment of local personnel, etc., which have presented major problems for SERS's own operation, and because rural roads planning capacity by BPP is weak at present, since the counterpart trained by UNDP technical assistants has been promoted to Director of New Works and Coordinator for the Association's Fourth Highway Project. 4.07 The project provided training to a significant number of equip- ment operators and contributed to the training of 26 road section Chiefs who followed theoretical and practical sessions at the Publics Works Training Center in Ouagadougou. A further two road section chiefs, four mechanics and two topographers were trained at the CERFER in Lome. All of the trainees were subsequently employed by SERS. 4.08 The project's contribution to the development of local contrac- tors did not materialize (cf. para 3.11 above). But it contributed to the development of local consulting capacity. A local firm, after work- ing in association with a foreign consultant on the socio-economic impact study, subsequently carried out an inquiry for the project into the potential for community based road maintenance. Again, DPW has not dis- cussed the findings of the study which indicate good potential for invol- ving local communities in maintenance, essentially because officials in charge are uneasy and unenthusiastic about moving away from.present methods which they control directly. V. ECONOMIC RE-EVALUATION 5.01 At appraisal, the expected economic rate of return based on vehicle operating cost savings was estimated at between 10-18% for both the construction and maintenance components. New roads included into the construction program in later years were evaluated in terms of incremen- tal producer surplus in the road influence areas with estimated returns ranging from 12-25%. In the absence of any traffic counts on the project roads, or of ORD-level data on agricultural production after 1977/78, there is no basis for up-dating rate of return estimates realistically. Under the follow-on project a regular schedule of traffic counts is to be set up, and BPP should undertake annual up-dates of ORD agricultural statistics and survey local prices for main staples marketed to establish medium term trends of marketable and marketed agriculture products in the regions where roads are being improved as part of ongoing rural develop- ment projects. Rough estimates (see Table 6) indicate that actual costs of road improvements have increased by some 10% over those estimated at - 36 - appraisal for equivalent standard Type A roads (US$8,600/km). Unit vehicle operating cost savings have increased by about 50% but accrue on only about 60% of the estimated road length. It can thus be concluded that the road investments yield a satisfactory rate of return of 10% or more as at appraisal with an expected average traffic as low as 10-12 vehicles per day. Returns to road maintenance are likely to exceed 20%. VI. ASSOCIATION PERFORMANCE 6.01 The Upper Volta Rural Roads Project has been treated from prepa- ration through completion by the Bank's Regional Mission in West Africa (RMWA). It was the second in a series of rural roads projects in West Africa Region that attempted to establish a single major focus for rural road planning, construction and maintenance. The project was timely in the Upper Volta context: it had full Government support from inception and was allocated experienced personnel and the requisite counterpart funds. The flexibility provided at appraisal for later-year construction programming proved necessary to keep the project responsive to changing rural development priorities over the eight year project gestation and execution period. Cost estimates included substantial price contingen- cies that would have been generous, had the project been completed over the three-year period foreseen at appraisal. 6.02 Supervision missions averaged between two and four annually. With the possible exception of a one year gap between missions in 1978, there is no reason to believe that increased supervision intensity would have affected project performance. During the apparent year gap between missions in 1980, preparation missions for a planned second feeder roads project (later incorporated in the Fourth Highway Project) took place, during which minor problems were easily resolved. Frequent supervision visits during the last project year however had but limited impact on a deteriorating situation that effected many government institutions in the country. In all, the Association contributed significantly to the crea- tion of a competent rural roads organization in DPW, although performance has of late declined due to a number of mostly extreneous factors (see para. 4.03 above). VII. CONCLUSIONS 7.01 The Rural Road Project was well conceived to support Upper Volta's agricultural development strategy and, with its combined emphasis on both upgrading and regular maintenance of the rural road network, timely in its objective of making the competent technical ministry res- ponsible for the upkeep of infrastructure improvements financed under - 37 - various earlier and ongoing regional development projects, that would otherwise have faced rapid deterioration. Fully supported by Government, the project had by early 1980 realized most of its institutional objec- tives as well as rates of work output close to appraisal estimate; how- ever, performance declined sharply in the last year of project execution, due to both institutional and political upsets. Construction and maintenance outputs of the project took two years longer than estimated to produce and in terms of kilometrage constructed and maintained amounted to 61% and 88% respectively of outputs foreseen in the indicative program defined at appraisal; however, since construction was to higher standards than originally foreseen, it amounted to over 95% of the volume at appraisal. 7.02 As regards institution building the project points up two les- sons: the Association tends (a) to underestimate the time necessary for a new institution to reach operating efficiency, and (b) is too ready to promote organizational change rather than consolidate and improve opera- tion of existing organizations. Thus, the project appraisal assumed that the new SERS could reach normal construction outputs of 25 km/month after a six month start-up period; actually, it took about two years to achieve this rate of output. The appraisal was over-optimistic in its assessment of (i) the serviceability of existing equipment to be taken over by the project and the willingness of previous users to part with it on a given date; (ii) the speed with which competent head office/field/technical assistance personnel can be found and constituted into a working team and (iii) the time required for new local institutions to learn coping with Association procurement/disbursement procedures effectively. Then, less than a year after a team of local counterparts trained by the project had started working satisfactorily, Association-backed decentralization pro- posals for DPW disrupted arrangements and performance declined. 7.03 Institutional progress was most difficult to achieve on non- engineering aspects of project implementation. Preoccupied with the practical problems of finding competent technical personnel and keeping it regularly paid, and of having a minimum of construction and main- tenance equipment operational, DPW management tends to pay insufficient attention to the improvement of road planning and economic evaluation procedures. Despite long-term technical assistance to BPP, basic trans- port planning data (number and type of vehicles, traffic counts, volume and flow of agricultural goods) remain deficient or non-existent; re- sults of studies undertaken by consultants are ill-assimilated or un- used. Activities involving non-DPW personnel such as ORD directors for defining road priorities, or local community representatives for organiz- ing labor-based maintenance -- are difficult to initiate and sustained only at Association insistence. - 38 - Table 1 UPPER VOLTA RURAL ROADS PROJECT (CR. 579-UV) PROJECT COXPLETION REPORT Design Standards: At A2praisal and Actual. Appraisal Report Actual Designation Type A Type B TyDe -C Clearin3 12 m 10 m 12 m Roaday 7 m 6 m 6 m Gravelling width 5-6 a (a) 4.5-5 (a) 5 a (b) Gravelling depth 0.15 a 0.12 m. 0.20 a Gradient 10% 12% 10% Design Speed 60 ka/h 40 km/h 60 1n/h (a)' Spot improvements. (b) Continuous gravelling. RURAL ROADS PROJECT (CR. 579-UY) PROJECT COMPLETION RETORT Rural Road -Constructed by Region and Year Road N Road Secions by Region BPP Total Constructed Total % of Programmed 1977 -197 197 18r- rT11T Constructed I Structures F Km Km Km Km Km Km Completad COME (Banfora) 139 - 43 88 8 - 139 R 38 Studon Naguera (1) 35 - - - IO0 - (N7-YSE) Banfors - Hangodara (4) 104 - 8 8 8 - 104 100 Z HAUTS BASSINS (Bobo'Dioulasse! 163 40 41 - 28 20 129 R 35 Orodara-Kourouma-NDorola (1) 101 40 41 - 81 100 2 R 3d Bobo-Dioutasso - Banzon 62 20 48 602 VOLTA NOIRE (Dddougou) 332 - 71 107 46 39 263 R 16 Teugan - Nohoro (4) 85 71 14 85 100 x R 20-21 Solenzo - Tansila (4) 55 5' 55 ICO 2 D 18. R 2d Solenzo - Sanaba - Balavd (4) 84 38 1. 84 100 2 R 20 Sanaba - Kodougou (4) 34 34 34 100 z D 14 Boromo - Safand (2) 74 5 5 I00O CENTRE - OUEST (Koudougou) 137 - - 41 23 96 - (R 24) Fara - Ouesa (3) 65 47 I8 - 65 90 z _(N14E112) Koudougou - Kindt (5) 72 5 26 31 95 CENTRE (Ouagadougou) 35 - - 35 - 35 -(01548W0115) Ktombltrl-Gana-Toesse (4) 35 35 35 100% CENTRE - EST (Koupila) 24 24 - - - - 24 100 x -(N17-NE) Hoaga - Bissiga (3) 24 24 24 50 EST - (Fada N'Gourma) 50 40 10 - - - 50 -(Nh 91SW) * Fada N'Gourma - Comin Yanga (2) 50 40 10 50 50 % TOTAL 880 104 165 242 3.40 85 736 R = Regional Road (1) = First year program--firmly established D = Departmental Road (2) = Indicative program N = National Road (3) - Improvements initiated under RDF Project (R) - Off Regional Road x in direction indicated (4) = Newly evaluated roads included at ORD request (5) = Improvement priority for Koudougou Rural Develop- ment Project &’豐！&! UPrFR VOLTA RURAL ROADS PRIJECT (Cr. 579-11Y) P ECT COMPIXTION REPORT Not of Regional and Rural Rands Naintained bi_11,,ALS1 and Year ToT KM H A 1 N T A I N E D Total Grand Road NO Road Section by Region Lou th 1977 1978 1979 1980 1981. Maintained TOTAL H N It M H N N 11 N H t: CENTRE (Oussadausou) 659 211 275 16 1215 195 8 570 430 --14-00Q Ouagadougou - Kongousal 115 its its Its 1 230 Its D 2 Kongousat Dj1ba, 55 S5 35 -(D 2) Kongousel Bourzanga 45 45 45 -(D 2) Rollo - Ousse 20 20 20 N3 Ouagadougou - Keys 105 105 105 N6 Ouagadougou - Boulbi. 24 24 24 -(NI) Ouagadougou - KomsLISs -CN5) Dagouma Bazega 27 27 27 -05) R N 5 Gutrga it it 11 -(N5) R N 5 Hangs 23 23 23 D5 Man$& Zabre 75 75 75 75 75 DS Hangs NobLli R N 1 25 25 25 -(D5-RlI) Hangs Kaibo 12 12 12 -(D5-RII) R N 5 Binde, Katbo 30 a 30 30 24 to of It a R N 5 Rinde Nobare 45 45 45 R N 5 Gaongo to is is R N 5 Bazaga 19 14 19 33 R N 5 Timbane CENTRE EST Mupela) 236 _j g 20 = ..I. _L76 0 41 88 57 493 D 9 Tenkodogo - Garango 20 20 20 20 '0J 20 80 D 9 1. - 20 D 9 Garango - KaLbo 68 69 68 30 30 272 11 11 68 TankodoSo - Badego 7 7 N 11 Tenkodog Ouarad a .60 60 60 -(N 17) Moaga - 3:;.iFta '2Y 24 24 24 -(N 4) Bisaign - Zanre 30 30 30 -(N 4) Urge - Weguat 27 27 27 EST (Fada-Aga-uM) 37 .!7 247 27 274 311, -(N17-R6) Quargaye - Comin Yanga 27 27 27 27 27 54 N 4 Veda - Kantchari ISO 130 150 N 4 Kantcharl. - Bouttou - F. Niger 70 70 70 GRAND TOTAL 3,287 52 329 20 689 605 902 422 1,652 209 640 1,308 4,212 S,52o .......... ........... ........... ........... ........... ............ ....... Included in indicative program for upgrading 2/ Roads( re)-const rue t ad under the project. Bougouriba road& were maintained by the Agriculture Project Brigade dur,ng 1977-80. - 42 - Table 4 UPPER VOLTA RURAL ROADS PROJECT (CR. 579-UV) PROJECT COMPLETION REPORT Road Impact Study: Number of Zones, Road Sections Villages and Markets Selected for Study Sections Ka Standard (i) Banfora area Banfora-Sindou 50 P A 31 (10 villages incl. 2 controls Sindou-Baguera 33 P A A2 3 markets) (ii) Orodara area Orodara-Samoragouan 50 P A A (9 villages incl. 0 control 3 markets) (iii) Koudougou area Samendeni-Banouale 51 P A3 A (15 villages, incl. 2 controls Banouale-Lahirasso 20 P 0 5 markets) Labirasso-Kouka 34 P A B Banouale-Bakonorisso 16 P 0 Koundougou-Solenzo 70 R T M4 (iv) Oronkua area Oronkua-Fing-Founzan 53 P A A -(8 villages, incl. 1 control 3 markets) (v) Nako area Nako-Bouroum Bouroum 26 P A B (20 villages, incl. 3 controls Nako-Bomoi 17 P A B 5 markets Nako-De-oola-Bomoi 52 P A A (vi) Kongoussi area Kongoussi-Djibo 92 R T M (14 villages, incl. 1 control 5 markets) (vii) Fada N'Gourma area rada N'gourma-Comin 50 P A A (3 villages, incl. 1 control Yanga 2 markets) TOTAL SAMPLE: 7 Zones, 14 Road Sections, 20 Mrkets, 80 Villages, 320 Farahouseholds. 1 Piste Amelioree B - Track improved to B standard (of Table 1) 2 Piste Amelioree A = Track improved to A standard (of Table 1) 3 Piste Ordinaire - Unimproved Track Route en Terre Moderne = gravel road N.B: Note that only one of the Study Roads (Fada N'Gourma-Comin Yanga) was improved under the project. UPPER VOLTA RURAL ROADS PROJECT (CR. 579-UV) PROJECT COMPLETION REPORT Appraisal Estimate and Actual Project Costs (Current Prices, Without Taxes) Appraisal Estimate Actual Costs Foreiln Total Cost Item Local Foreign Total 1/ Local Foreign Total 2/ as % of as X of CFAF million US '000 U9$'0 CFAF million US_ '000 US=000i Total Cost Appraisal Estimate A. EQUIPMENT + MATERIALS 15.2 3,003.1 3,)70.7 36.0 3,246.3 3,396.5 95.5 110.7 Equipment 10.0 2,133.3 2,177.8 25.0 2,439.6 2,544.0 117.0 Materials + Supplies 5.2 869.8 892.9 11.0 806.7 852.5 95.5 B. SPARE PARTS + FUEL 99.1 2,514.7 2,955.1 35.5 2,845.4 2,993.3 94.9 101.3 Spare Parts. not available 16.3 1,315.4 1,383.3 Fuel 19.2 1,530.0 1,610.0 C. STAFF 165.3 28.0 762.6 370.8 - 1,545.0 - 202.6 Head Office 21.7 28.0 124.4 92.6 3/ 385.8 310.1 Field (Brigades) 143.6 - 638.2 278.2 1,159.2 181.6 W D. TECHNICAL ASSISTANCE 11.9 320.9 373.8 11.5 136.7 184.6 74.0 49.4 E. WORKS CONTRACTED 117.5 524.4 1,046.6 - - - - F. STUDIES - 257.8 257.8 58.8 205.0 450.0 45.6 174.6 TOTAL FINANCED BY PROJECT 409.0 6,648.9 8,466.6 512.6 6,433.3 8,569.3 75.0 101.3 ammamm.m. mmmm.mm.....mu.... -----------f---*m........ .m Financed by IDA 191.1 6,648.9 7,500.0 255.7 6,433.3 7,494.7 PRO HEM Technical Assistance Financed under Credit 706-UV 10.1 274.8 317.6 Estimated Cost of Head Office Personnel Financed by Government 25.0 Construction of SERS Headquarters--1978-1980 260.2 Furnishing of SERS Headquarters--1980 24.4 Construction of SERS Workshop--1981 48.2 1/ Estimated ecchange rate US$1 - CFAF 225 2/ Average actual exchange rate US$1 - CFAF 240. Rates fluctuated between CFAF 200-340 per US$ during 1976-1982; hence the difference in dollar amounts spent per expenditure category to those shown in Table 8. 3/ Includes CFAF 9 million for training. -44- Table 6 UPPER VOLTA RURAL ROAD PROJECT (CR. 579-UV) PROJECT COMPLETION REPORT Estimation of Average per Kilometer Cost of Roads Improved under the Project (Current prices, net of taxes in FCAF million) Residual Value 1/ Item Total Cost and Stocks Equipment 610.5 200.0 Spare Parts and Tyres 332.0 45.0 Fuel and Lubricants 386.4 - Materials and Supplies 204.6 20.0 Salaries 370.8 - TOTAL 1,904.3 265.0 Cost of Road Maintenance 193.2 (5,520 km Ld 35,000 FCFAkm) Total Cost of Road Improvements 1,711.1 Kilometers Improved 736.0 Cost per km improved: 2.325 - US$9,680 I/ The residual value of equipment and stocks have not been taken into account to compensate for the value of equipment inherited from the two agriculture projects. - 45 - Table 7 UPPER VOLTA RURAL ROADS PROJECT (CR. 579-UV) PROJECT COMPLETION REPORT Estimated and Actual Schedule of Disbursements (FY 1977-1982) IBRD/IDA Cumulative Disbursements at End of Quarter Fiscal Year (US$ '000) and Quarter Estimated (E) Actual (A) A as Z of E 1977 1 100 2 500 3 2,700 4 3,600 1,000 28 1978 1 4,800 1.000 2 5,000 1,100 3 5,500 1,400 4 6,000 2,000 33 1979 1 6,600 2,400 2 7,100 2,800 3 7,300 3,200 4 7,500 3,300 44 1980 1 3,500 2 3,700 3 4,000 4 4,500 60 1981 1 4,700 2 5,200 3 5,700 4 6,000 80 1982 1 6,500 2 6,500 3 6,600 4 7,500 1/ 100 1/ A balance of US$5,813 was cancelled. Source: Bank Government and mission reports. - 46 - Table 8 UPPER VOLTA RURAL ROAD PROJECT (CR. 579-UV) Estimated and Actual Allocation of Credit Proceeds (USS Equivalent) Credit Reallocation Credit Category Agreement July 1979 Actual (1) Equipment Spare Parts, Materials and Supplies 4,100,000 5,100,000 6,053,056 (2) Consulting Services and Technical Assistance 500,000 695,000 651,078 (3) Operating Expenditure for Better ment and Maintenance Work under the Project 250,000 560,000 790,053 (4) Civil Works 550,000 550,000 - (5) Unallocated 2,100,000 595,000 TOTAL 7,500,000 7,500,000 7,494,187 祖 IBED11504(PC å OCTOSER 1962 k- Ta Ca~ DORI 0 SEf^ EGA 1 1. K Räu 0 p~ 90GAND9 4. 0 GOULSA INDARE AGAGMQUGO LA FADA-tTGOURAMA q8 DIAPAGA 0 mm NIO 1;. i Gu, Saltengå T TEN ew FN~ i oimii Ko~Yanga .0i RS, Duarpwc oenge, Påmi. B E N l N 1,0 11 Ti- rabie )m 4.- H A N T 0 G 0 EARTH RURAI. ROAD TYPE A TO BE IMPROVED OR MAINTAINED UNDER THE PROJECT ROADS CON- AD TYPE B TO BE INPROVED JER THE EARTH RURAL RO OR MAINTAINED UNDER THE PROJECT ROADS MAIN- i THE PRojuvr T RACKS TO 8 E MAI NTAJ NE D TAINED BY z--ý PREVIOUS BANK PROJECT 3RIGADE THE ROAD NETWORK IN THE SHADEO AREA 15 UNDER STUDY FINANCED BY THE BOUGOURIELA PROJECT CREDIT NOL Agrl-UV. DEPARTEMENT BOUNDARIES A CERCLE. souNDARIES IORD EXISTING PAVED ROADS e i G C) woPK PmVED ROADS UNDERSTUDY ýc -e RED. G0,4 UPPER AMIES SHOWN. GRAVEL POADS 0 50 60 70 INTERNATIONAL BOUNDARIES lvoR;, r I I I . ta ERS COASTtir ORDiREGIONAL CENTERS ,I